MEGA-Bench - Application: Information Extraction — leaderboard

Metric: Average Score (%). Source: huggingface.co. 44 models tracked.

Top models

#ModelScore
1Gemini 2.5 Pro72.6
2GPT-4o70.6
3Claude 3.5 Sonnet (20240620)66.6
4Claude 3.5 Sonnet (20241022)65.9
5Gemini 2.0 Flash (Preview)64
6GPT-4o Mini56.7
7Qwen 2 VL 72B56.6
8InternVL3-78B55.5
9Gemini 1.5 Pro (002)54.2
10InternVL2.5-78B54
11InternVL3-38B52.5
12InternVL3-14B49.5
13Gemma 3 27B (IT)49.3
14Gemini 1.5 Flash (002)48.2
15InternVL3-8B44.8

Interactive version: theaggregate.ai/benchmark?slug=mega-bench-application-information-extraction · How the rankings work · Data refreshed daily, snapshot 2026-07-22.